{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T23:00:17Z","timestamp":1790290817226,"version":"4.1.0"},"reference-count":54,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["J Med Internet Res"],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Physician documentation requirements are a known contributor to clinician burnout, with the manual creation of brief hospital course (BHC) summaries being particularly time-consuming. Automating BHC summarization may mitigate this workload and reduce documentation errors. However, current natural language processing (NLP) methods are often limited to single-document inputs, and large language models (LLMs) face privacy and deployment challenges. Furthermore, existing methods often require manual information extraction and struggle to maintain temporal accuracy.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>We developed and evaluated LiteMedDoc, a lightweight, locally deployable LLM-based framework to automatically generate BHC summaries without model fine-tuning. Our objective was to determine whether clinically useful clinical summaries could be generated under strict privacy and resource constraints, making automated summarization feasible in real-world hospital environments.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>LiteMedDoc is a modular pipeline built upon an 8-billion-parameter open-source LLM (Llama 3.1). It features 3 specialized modules: a static dynamic information hierarchy module to condense multisource inputs and structure clinical events chronologically; a similar document retrieval-augmented generation module that retrieves contextually relevant prior case summaries; and a self-adaptive feedback optimization module used offline for prompt optimization. All processing was performed locally without any model fine-tuning. We evaluated the framework on a retrospective cohort of 4538 coronary artery bypass grafting (CABG) surgery cases from the Medical Information Mart for Intensive Care (MIMIC)-IV database. A held-out test set of 403 cases was used to generate BHC summaries. The model-generated summaries were compared to reference BHCs using 8 standard NLP metrics covering lexical overlap (BLEU-4 [Bilingual Evaluation Understudy-4 gram] and ROUGE [Recall-Oriented Understudy for Gisting Evaluation]), semantic similarity (BERTScore and METEOR [Metric for Evaluation of Translation With Explicit Ordering]), and clinical relevance (AlignScore [Alignment Score] and MEDCON [Medical Concept Overlap]). Additionally, 15 cardiac surgeons conducted a clinical evaluation of a sample of model-generated summaries, rating them on completeness, correctness, readability, conciseness, and global quality using a 5-point Likert scale.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Without any model training, LiteMedDoc achieved strong performance across individual automated metrics, nearly matching a fine-tuned model and exceeding a 70-billion-parameter model on all metrics. Additionally, in a within-database cross-domain evaluation on lobectomy cases, the framework maintained encouraging performance after prompt adaptation and outperformed both the base model and the CABG-fine-tuned model. Surgeons rated the AI-generated summaries above the prespecified acceptability threshold across all domains (mean scores \u22653.0), specifically praising their structure and conciseness.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>By integrating 3 specialized modules, the proposed framework offers a practical, locally deployable solution for clinician-in-the-loop BHC draft generation under privacy and resource constraints, and holds promise for improving documentation efficiency and enhancing information continuity during care transitions.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.2196\/90870","type":"journal-article","created":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T22:30:08Z","timestamp":1790289008000},"page":"e90870","source":"Crossref","is-referenced-by-count":0,"title":["Automated Brief Hospital Course Summarization in Cardiac Surgery Using a Lightweight Large Language Model\u2013Based Framework: Development and Evaluation Study on the Medical Information Mart for Intensive Care-IV"],"prefix":"10.2196","volume":"28","author":[{"ORCID":"https:\/\/orcid.org\/0009-0007-6433-6098","authenticated-orcid":false,"given":"Xiaoyuan","family":"Gao","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4843-2673","authenticated-orcid":false,"given":"Yang","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-2356-8466","authenticated-orcid":false,"given":"Zixing","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5345-0070","authenticated-orcid":false,"given":"Jing","family":"Yuan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-9714-3548","authenticated-orcid":false,"given":"Shengkang","family":"Huang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9260-188X","authenticated-orcid":false,"given":"Xu-Yao","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1793-1162","authenticated-orcid":false,"given":"Zhaohong","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-4290-0931","authenticated-orcid":false,"given":"Yun","family":"Xing","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-6463-3079","authenticated-orcid":false,"given":"Yiyang","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9521-640X","authenticated-orcid":false,"given":"Xintong","family":"Wu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-5736-3793","authenticated-orcid":false,"given":"Zhan","family":"Hu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2127-9624","authenticated-orcid":false,"given":"Wei","family":"Zhao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,9,24]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1016\/j.mayocp.2016.05.007"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jad.2022.12.153"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijnurstu.2022.104385"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.7326\/M16-0961"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.2196\/55247"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1001\/jama.297.8.831"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.2196\/57721"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1001\/jamainternmed.2025.0821"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1016\/j.sapharm.2025.03.062"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.2196\/48568"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1016\/j.jbi.2023.104358"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1016\/j.pec.2013.04.019"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1093\/jamia\/ocv032"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1007\/s11606-025-09740-y"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pdig.0000158"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1080\/08820538.2023.2209166"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.2196\/68998"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijmedinf.2024.105654"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1016\/j.annemergmed.2024.01.039"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/j.lanwpc.2024.101211"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1016\/j.asoc.2024.111531"},{"key":"ref22","unstructured":"XuJDischarge me: BioNLP ACL-24 shared task on streamlining discharge documentationPhysionet.org2026-04-19https:\/\/physionet.org\/content\/discharge-me\/1.2\/"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1016\/S2589-7500(23)00021-3"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-025-01618-7"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-022-01899-x"},{"key":"ref26","doi-asserted-by":"crossref","unstructured":"GuoRFarnanGMcLaughlinNDevereuxBQUB-Cirdan at \u201cDischarge Me!\u201d: zero shot discharge letter generation by open-source LLM2024Proceedings of the 23rd Workshop on Biomedical Natural Language Processing2024 August 16Bangkok, Thailand664674","DOI":"10.18653\/v1\/2024.bionlp-1.58"},{"key":"ref27","author":"Dubey, A","year":"2024","journal-title":"arXiv. Preprint posted online on July 31, 2024"},{"key":"ref28","unstructured":"LewisPPerezEPiktusARetrieval-augmented generation for knowledge-intensive NLP tasks202034th Conference on Neural Information Processing Systems (NeurIPS 2020)2020 December 6-12Vancouver, Canada"},{"key":"ref29","unstructured":"WangWWeiFDongLBaoHYangNZhouMMINILM: deep self-attention distillation for task-agnostic compression of pre-trained transformers202034th Conference on Neural Information Processing Systems (NeurIPS 2020)2020 December 6-12Vancouver, Canada"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1023\/a:1010933404324"},{"key":"ref31","author":"Jiang, AQ","year":"2023","journal-title":"arXiv. Preprint posted online on October 10, 2023"},{"key":"ref32","author":"Yang, A","year":"2025","journal-title":"arXiv. Preprint posted online on May 14, 2025"},{"key":"ref33","unstructured":"GPT-4o mini: advancing cost-efficient intelligenceOpenAI20242026-09-04https:\/\/openai.com\/index\/gpt-4o-mini-advancing-cost-efficient-intelligence\/"},{"key":"ref34","doi-asserted-by":"crossref","unstructured":"PapineniKRoukosSWardTZhuWJBLEU: a method for automatic evaluation of machine translation2002ACL '02: Proceedings of the 40th Annual Meeting on Association for Computational Linguistics2002 July 7-12Philadelphia Pennsylvania311318","DOI":"10.3115\/1073083.1073135"},{"key":"ref35","unstructured":"LinCYROUGE: a package for automatic evaluation of summaries2004Text Summarization Branches Out2004 July 25\u201326Barcelona, Spain7481"},{"key":"ref36","unstructured":"ZhangTKishoreVWuFWeinbergerKBERTScore: evaluating text generation with BERTOpenReview20202026-09-04https:\/\/openreview.net\/pdf?id=SkeHuCVFDr"},{"key":"ref37","unstructured":"BanerjeeSLavieAMETEOR: an automatic metric for MT evaluation with improved correlation with human judgments2005Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and\/or Summarization2005 June 29Ann Arbor, Michigan6572"},{"key":"ref38","doi-asserted-by":"crossref","unstructured":"ZhaYYangYLiRHuZALIGNSCORE: evaluating factual consistency with a unified alignment function2023Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)2023 July 9-14Toronto, Canada1132811348","DOI":"10.18653\/v1\/2023.acl-long.634"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-023-02487-3"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2024.48723"},{"key":"ref41","doi-asserted-by":"crossref","unstructured":"DammHPakullTEry\u0131lmazBWisPerMed at \u201cDischarge Me!\u201d: advancing text generation in healthcare with large language models, dynamic expert selection, and priming techniques on MIMIC-IV2024Proceedings of the 23rd Workshop on Biomedical Natural Language Processing2024 August 16Bangkok, Thailand105121","DOI":"10.18653\/v1\/2024.bionlp-1.9"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2024.48729"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2017.8444"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2025.26339"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.2196\/55318"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.2196\/63445"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06291-2"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.2196\/19810"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.2196\/49724"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1056\/aioa2300068"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.2196\/80557"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.12788\/fp.0589"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-024-01083-y"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2025.26350"}],"container-title":["Journal of Medical Internet Research"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jmir.org\/2026\/1\/e90870","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T22:30:15Z","timestamp":1790289015000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jmir.org\/2026\/1\/e90870"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,24]]},"references-count":54,"URL":"https:\/\/doi.org\/10.2196\/90870","relation":{"has-preprint":[{"id-type":"doi","id":"10.2196\/preprints.90870","asserted-by":"object"}]},"ISSN":["1438-8871"],"issn-type":[{"value":"1438-8871","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,24]]}}}